There exist situations where an upgraded OS is simply not compatible with good old software. These systems may be connected to legacy hardware that cannot be easily upgraded or replaced. These systems have worked for decades and are not connected to the network. These systems often predate monumentally stupid ideas like automatic updates, embedded spyware, and software subscriptions. Rather than blindly telling everyone to upgrade it would be more useful to help maintain legacy compatibility of retro systems that you may not even understand the reasons for existing. Thank you! | [reply] |
Rather than blindly telling everyone to upgrade it would be more useful to help maintain legacy compatibility of retro systems that you may not even understand the reasons for existing.
The onus of maintenance is on the people who need to keep using old, decrepit, obsolete software, including OSen. Not on anyone else. If that is too much, such an enterprise ought to hire people who can. Yes, tough luck🤷♂️
| [reply] |
not compatible with good old software...legacy hardware
++ for bringing that up. And hardware doesn't just mean CPU and HDD and RAM -- it can be other proprietary hardware connected-to-and-run-by the good old software. So "update OS" or "update perl" isn't always an option. (That's has been one of my situations at $work.)
maintain legacy compatibility of retro systems
On the other hand, I disagree with this. The developers of perl and developers of CPAN modules shouldn't be beholden to those of us with legacy needs. Perl makes available old copies of the language, so that's not a problem. CPAN keeps old versions of modules, which you can download (the common CPAN clients even allow you choosing version and/or URL for a given module's distro) ... and even if the author has deleted ancient versions of their modules from active CPAN, it's still on the BackPAN. So if you have a set of perl-version-plus-modules that worked a decade ago, that same set of perl-version-plus-modules is still available to you.
| [reply] |
| [reply] |